GRATINGS STIMULI DO BIAS OUR CONCEPTS ON CORTICAL GAMMA SYNCHRONIZATION: A STUDY IN CAPUCHIN MONKEY V1
gamma, synchronization, V1, visual perception, capuchin monkey.
Cortical gamma oscillations (30 - 90 Hz) have been implicated in various cognitive processes, such as visual binding and attention. It has been hypothesized that the precise phase relationship of oscillatory signals are important for stimulus encoding, and ultimately the control of information flow in the brain. So far, most evidence in support of this hypothesis are based on studies that used artificial, simplified stimuli, such as moving gratings and bars. Recently, studies using natural images led to conflicting conclusions. In a paradigm in humans that required prolonged fixation, ECoG signals showed gamma for grating stimuli but not for static images or pink noise (Hermes et al., 2015). In V1 of capuchin monkeys, spiking activity exhibited strong beta but no gamma during free viewing of static images (Ito et al., 2011). Conversely, analysis of ECoG signals in the early visual cortex of macaque monkeys revealed strong gamma components for natural scenes (Brunet et al., 2015). In the present study, we aim to clarify these discrepancies using a paradigm that that allowed direct comparisons between fixation vs. free viewing conditions, for both simplified stimuli (moving and static gratings) and natural scenes (static and moving images). Within a trial, the monkeys were required to fixate for 2 seconds (FIX epoch), after which they could freely inspect the visual stimulus for another 2 to 4 seconds (FREE epoch). In some experimental sessions, the monkeys were also exposed to real world scenes, such as viewing of other monkeys, humans or real objects. This last condition, albeit not as well controlled, provided a means of assessing visual responses in truly naturalistic conditions.
Recordings of spiking activity and local field potentials (LFPs) were made by means of 2-4 quartz-insulated electrodes placed at the central and peripheral representation of V1. The electrodes could be moved independently at different layer positions in the cortex. At the beginning of each recording session, receptive field (RF) maps were obtained using a single drifting bar. Thereafter, orientation selectivity was evaluated by means of large, high-contrast gratings. Gratings (optimal and non-optimal directions) and natural scene pictures (moving at the optimal and non-optimal directions) were then presented 10 to 20 times in a randomized, interleaved order. During the FIX epoch, stimuli were presented through an aperture mask (~10°) centered on the RFs, while during the FREE epoch they covered most of the screen display area.
Our results provide first a general assessment of gamma responses in capuchin V1 (N= 3 monkeys). In the capuchin, oscillations can be as strong as those seen in macaque monkeys or in humans. Gamma is strongly dependent on stimulus features, such as orientation, size and speed. Notably, in our data we observed the same eccentricity-dependent effect on oscillation frequency described previously for the macaque (Lima et al., 2010).
Direct comparisons made for gratings and natural scenes revealed that gamma responses were invariably stronger for gratings, and relatively weaker or absent for natural scenes (N= 100 recording sites, 2 monkeys). This general trend was seen both for the FIX and the FREE epochs. For a gratings moving at the optimal direction, driving the cells uninterruptedly, gamma was distinctly strong. Similarly, the free viewing of a large optimal moving gratings led to strong gamma, despite of intervening saccades and eye blinks. For the responses to natural scenes, however, gamma was present only occasionally. Brief gamma responses could be seen in some recordings, generally when the image moved at the preferred direction and during the FIX epoch (6 of 50 recording sites). It is possible that in these cases gamma appeared only for responses to contours in the image that precisely matched the preferences of the cells. Our results for more naturalistic conditions were, however, surprisingly negative. Free viewing of natural scenes yielded flat correlograms both for moving and static images. Similarly, the analysis of responses during free observation of real scenes revealed no signatures of gamma.
Our results suggest that strong, narrow-band, gamma responses in the visual cortex is primarily associated with the selective activation of cell populations. Since it is known that in the visual cortex columns sharing the same properties are preferentially connected (Schmidt et al., 1997), gamma may be considered a „special case“ dynamics largely shaped by cortical connectivity. Moreover, gamma oscillation frequency seems to depend on the cortical map, since it strongly depends on RF eccentricity. This new finding in a new-world monkey reinforces our hypothesis that the functional architecture imposes important constraints on coherence dynamics in the cortex. Overall, the present results weaken the notion that gamma works as a flexible control mechanism, necessary for visual processing in natural conditions.
Our results suggest that strong, narrow-band, gamma responses in the visual cortex is primarily associated with the selective activation of cell populations. Since it is known that in the visual cortex columns sharing the same properties are preferentially connected (Schmidt et al., 1997), gamma may be considered a “special case” dynamics largely shaped by cortical connectivity. Moreover, gamma oscillation frequency seems to depend on the cortical map, since it strongly depends on RF eccentricity. This new finding in a new-world monkey reinforces our hypothesis that the functional architecture imposes important constraints on coherence dynamics in the cortex. Overall, the present results weaken the notion that gamma works as a flexible control mechanism, necessary for visual processing in natural conditions.
All experimental procedures were previously approved by the Comitê de Ética da Universidade Federal do Rio Grande do Norte (Protocolo N° 053/2012, CEUA-UFRN). Approval SISBIO N° 35425-4, 2016.